Partial and parallel pipeline processing in a deep question answering system

1. A method to reduce an amount of processing required to generate a response to a first case by a question answering system, the method comprising:

determining that a similarity score, of the first case relative to a second case, exceeds a similarity threshold;

identifying, by operation of one or more computer processors, a first feature of the second case having a first relevance score exceeding a relevance threshold, wherein the relevance score indicates that the first feature is relevant in generating a correct response to the second case;

determining, based on the similarity score and the first relevance score, that the first feature is relevant in generating a correct response to the first case;

identifying a first candidate answer for the first case that does not have the first feature; and

refraining from analyzing the first candidate answer in generating the response to the first case, thereby reducing the amount of processing of the question answering system.

2. The method of claim 1 , further comprising:

analyzing a second candidate answer in generating a response to the first case upon determining that the second candidate answer has the first feature.

3. The method of claim 1 , wherein refraining from analyzing the first candidate answer comprises:

refraining from performing natural language processing on the first candidate answer;

refraining from determining whether the first candidate answer has a second feature, wherein the question answering system has identified the second feature has a second relevance score exceeding the relevance threshold;

refraining from retrieving supporting evidence for the first candidate answer; and

refraining from scoring the supporting evidence for the first candidate answer.

4. The method of claim 1 , wherein the similarity score is computed based on a context of the first case matching a context of the second case.

5. The method of claim 1 , further comprising:

upon determining that: (i) a utilization of a resource of the question answering system does not exceed a utilization threshold, and (ii) a confidence score of a first response generated by refraining to analyze the first candidate answer exceeds a confidence threshold, generating a second response to the case, wherein the first candidate answer is analyzed in generating the second response to the case.

6. The method of claim 1 , wherein identifying the first feature of the second case as having the first relevance score exceeding the relevance threshold is based on a stored dependency of the second case to the first feature.

7. The method of claim 1 , wherein the first case comprises at least one question presented to the question answering system.

8. The method of claim 7 , wherein the first feature comprises at least one of: (i) a type, (ii) a subject matter, (iii) a variable, and (iv) a context of the at least one question.